The present invention relates to machine learning, and more specifically, generating a personalized viewing buffer from one or more inputs based on an established viewer-broadcaster relationship level.